Sunday, 10 November 2019

Governance or trust?

Since this series seems to be "Tom has a half-formed thought" let's look at governance vs trust. I regularly find myself having to find a comfortable place between the two and it's an interesting balancing act.

At one end of this spectrum, people rightly want to be trusted (and feel trusted, but that's a different post). When I ask "is X in a good place?" and the answer is "yes" that should be enough for me. My leadership team is highly skilled and I do trust them to do the right things.

However, at the other end of this spectrum, good governance tells me that I should have extensive metrics in place for the definition of "good" and that data flows straight to me so I don't have to ask this question, I can just look. I've already asked "give me these metrics and that will show me what you're doing".

I don't like either approach. I am not doing my job correctly if I don't have a clear understanding of what my technical leadership is doing and the decisions they are making. I also can't represent them effectively to the wider organisation without more detailed information. Ultimately, this hands-off approach leaves everyone at risk - while one can't really do the wrong thing when there are no parameters, with no specific questions being asked or answered, it's impossible to say someone did the right thing either. This is a serious problem when something goes wrong - and something always goes wrong eventually.

And I certainly don't want a deluge of data. I want to know that data is going to the right person and being acted upon. The right person isn't me - I have enough to do without interpreting unending raw data. This is why we have distributed leadership, and an empowered hierarchy. More importantly, the interpretation of the hard data is different in each area. I need to get to the story, and know the data is there if I need to dig into it.

Generally, I find that empowered autonomy is an excellent way to do things, however for the sake of everyone involved there has to be defined edges to that autonomy. That allows individuals to focus on their area, and gives me a structure for trust in the context of governance without it becoming intrusive. In practice, this means that I'm setting wide definitions and mostly looking at a traffic light rating system for ongoing reporting - giving very coarse-grained answers to my questions. This keeps a good balance between trust and governance. The raw data and interpretation of it is kept distributed where it belongs, but the specifics of each area means that questions are being asked and answered. The answers are trusted, but because the specifics are defined (can this be built? is the code linted? etc) there is enough of a shared understanding of what we mean by "yes" or "green". It also gives me a route to ask for the working, if there is need. It gives me what I need - a highlight of the areas that need discussion, and (hopefully) a sea of green for all the things that are fine.

Governance AND trust. I think.

This post is the second of five written in NaBloPoMo - the National Blogpost Month which, yes, is a thing. My plan is to write one post a weekend for the month of November. Due to some amazing planning, that means I have to write five posts rather than the four you might immediately expect. These posts will be a bit shorter than normal most likely and all of the posts will be tagged with (sigh) NaBloPoMo2019.

No comments: